hello,
I'm ph'd student. I have a statistical problem and you may help me.
I make my dissertation on the advertising efficiency of the internet. I
propose a conceptual model. I like to validate it with structural equation.
My problem is the following one : I've some dichotomous variables (click
the banner y/n). I use SAS to compute. I haven't procedure like Prelis in
SPSS so, (if I well read my statistical books) I need to compute a
.biserial corelation coeficient (no problem) and the associated covariance
matrix. This matrix is my main problem. I can't find the formula of this
matrix. I find the formula of "classical" covariance matrix but not for
.biserial matrix.
Could you help me please ?
Thank you for reading this message (and sorry for my english !)
****************************
Stéphane BOURLIATAUX-LAJOINIE
I.A.E de Poitiers
20 Rue G. Le Troubadour
BP 639 - F-86000 POITIERS
Tel : + 33 (0)5-49-45-44-99
E-mail : [log in to unmask]
****************************
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|